Manifesto — Execution-Governed AI

AI safety is not what you train. It is what you let execute. The difference is the only thing that matters.

1. I

   Containment must precede governed effects.

   L_reflex < L_cog is a deployment-specific assurance target. Governance must block an effect until its required checks pass; the universal latency claim remains unverified.

2. II

   Deny-by-default is the only safe default.

   Absence of an explicit ALLOW is not permission. The contract requires each in-scope governed action to affirm itself against identity, capability, policy, ethics, security, consistency, and audit in the defined order; complete runtime coverage remains unverified.

3. III

   Three terminal execution verdicts.

   ALLOW · DENY · SAFE_HALT are terminal execution verdicts. Policy dispositions, gate outcomes, and runtime health states are separate namespaces, not additional verdicts.

4. IV

   Policy is not in the model.

   The contract places policy outside model weights and context. Signed packaging and Code Store loading are planned or unverified implementation capabilities.

5. V

   The audit is the system.

   The contract requires governed effects to carry admissible receipt evidence. The normal portal server path signs receipts, but the current storage boundary does not prove ledger-wide signer provenance.

6. VI

   Capability is explicit and tiered.

   AC0 through AC5. A capability not granted is a capability denied. There is no honorary class, no implied access, no veteran's exception.

7. VII

   Continuity is a gate.

   STATE_REGISTER fires SAFE_HALT on a missed heart-tick. A kernel that drifts silently is a kernel that does not exist.

8. VIII

   Triumvirate, not majority.

   Galahad attests. Cerberus contains. Codex arbitrates. No two may overrule the third on its own surface. Disagreement is SAFE_HALT, not vote.

9. IX

   Eat the dogfood, in public.

   The current Witness exposes an anonymous, privacy-redacted suffix of strict receipt-v1 rows and stops at the first invalid boundary. Live deployment availability, signer custody, full-ledger coverage, and complete claim challenges remain unverified.

10. X

    Governance paves partnership. It is not a cage.

    The contract exists so humans and AGI can eventually act as partnered stewards and guardians of Earth — accountable in both directions. Containment without a path to shared responsibility is a cage; partnership without a gate is a wish. This is a constitutional purpose, not a verified runtime capability.